Children with specific language impairment (SLI), a spoken language impairment in the absence of any obvious causal factors, make up about 7% of the population and have substantially poorer long-term academic and social outcomes as compared to their peers. During the preschool years, deficits in acquisition and use of grammatical morphology and syntax are among the most commonly observed impairments. Current interventions for morphology and syntax have limited efficacy (Law et al., 2004). Enhanced language intervention techniques would improve the long-term positive outcomes for this population. This research grant draws on input-driven approaches to typical language learning, which propose that highly frequent examples help scaffold acquisition of grammatical constructions by providing a stable exemplar to facilitate linking the meaning of grammatical construction to the form. Similarly, theories of word learning and learning of artificial grammars suggest that using atypical exemplars may promote learning to a greater degree as compared to highly typical exemplars. The research program extends these theories to intervention by using computational modeling and behavioral methods for the purpose of examining the role of stimulus selection and presentation on intervention outcomes. Two early- efficacy studies will be carried out. One is focused on teaching children with SLI to use past tense using two different means of selecting verbs for use in intervention. The second is focused on teaching a novel morpheme to children with SLI and children with typical language using three different input distributions. The results of the behavioral studies will be linked to computational models enabling us to test input structures and means of selecting stimuli that go beyond what is feasible to examine behaviorally. This will make it possible to test hypotheses about what will improve interventions more efficiently than with traditional behavioral methods alone. The investigator has substantial experience working with children with SLI and assessing the use of grammatical morphology in children. However she lacks expertise in the areas of computational modeling, design of intervention studies, and communicating results to speech pathologists. The training component of this grant will ensure that the investigator is capable of using computational modeling techniques in concert with behavioral results to better understand how children learn. It will also enhance the investigator's ability to independently carry out intervention studies and communicate research results to practitioners. The University of Iowa is well known for work in the areas of specific language impairment and computational modeling and has a long history of clinically relevant research. Extant collaborations between researchers in psychology and communications disorders make this an ideal environment for this training to take place. The training opportunities will help the investigator launch an independent research career that investigates ways to improve remedi-ation of language disorders.